The way in which protein and DNA sequence similarities are computed can have significant impact on the reliability and sensitivity of database search methods. Proposed course: A version of the BLAST sequence database search application has been developed which allows explicit score matrices to be used with DNA sequence comparisons. The information content of DNA sequence alignments was calculated using the formalism developed by Altschul for protein B alignments. An application, BLASTX has been developed to translate nucleic acid sequences in all possible reading frames and search these conceptually translated protein sequences against protein sequence databases efficiently. The sensitivity of this algorithm to errors and uncertainties in the underlying query and database sequence data has been evaluated as has performance using raw expressed sequence tag (EST) sequence data.